Various types of video and electronic content present content that changes over time. Video, for example, typically consists of a series of frames that are played or otherwise displayed one after another. People, animals, automobiles, buildings, background, and other objects can appear on multiple frames of a video. For example, a person may appear to walk along a street from one side of a video to the other over the course of the video. More specifically, in the first frame of the video, the person may appear in a position at the far left of the displayed content, in the next frame the person may appear slightly to the right of its position in the prior frame, etc. Generally, the person may appear different in the various frames of the video. For example, as the person walks, the person's arms may swing back and forth and the person's body orientation may change.
For various reasons, tracking objects such as people, in recorded video and other electronic content presents various challenges. For example, it can be difficult to accurately identify an object in multiple frames of a video because the object may not be appear entirely the same and may appear in different positions in the various frames, as previously mentioned. Similarly, it can be difficult to determine the boundaries of an object. Generally, improved techniques for identifying, tracking, and using objects in videos and similar electronic content are needed.